Over the last year, the design team at Mill has logged hundreds of hours in commercial kitchens studying one of the least visible parts of restaurant operations: what happens to the trash.
The startup, which has raised more than $232 million since launching in 2020, is known for making a bin for home kitchens that dries and shrinks food waste. But companies like Whole Foods soon started asking for a version of their own. Of the 74 million tons of food that’s thrown out each year in the U.S., around a quarter comes from restaurants and grocery stores.
That’s both a major source of emissions and a significant cost for businesses with tight margins. But Mill is betting that better product design can help tackle the problem—and that unlike many climate challenges, reducing food waste can be relatively straightforward. “Food waste is one of those kinds of opportunities of massive scale impact that really is win-win across the board and is a solvable problem,” says Mill CEO and cofounder Matt Rogers, who pivoted to the challenge of waste after working on the original iPhone at Apple and cofounding Nest. “This is not fusion. This is not space travel. This is not putting food in the trash.”
The company’s new commercial product, under development now, tackles the problem from two angles. First, it turns kitchen scraps into a lightweight, non-smelly material that can be delivered to farms for use as chicken feed or composted. But the bins also use AI to analyze exactly what’s being thrown out—and help kitchens make changes to throw out less in the first place.
Inside Mill’s sprawling office south of San Francisco, multiple prototypes show how the device works. Like the consumer version, it starts by dehydrating food scraps. “The first thing we do is we pre-process food waste,” says cofounder Harry Tannenbaum.
Tannenbaum says 25% of the landfill is food and 80% of food is water—an insight that shaped the development of Mill. “If you take the water out of food, it gets really small,” he says. You can quickly turn it into something that looks like coffee grounds and smells like spices, and it’s easy to manage and a valuable resource.”
That approach convinced Whole Foods to reach out to Mill in 2024 as it looked for ways to cut food waste in half by 2030. “The major differentiator between Mill Commercial and other solutions was the on-site processing…the grind and dehydrate process allows us to reduce the volume of our food scraps by 80% on average, which will make it easier for our team members to handle, store, and transport,” says Cameron Crake, circularity lead for Amazon’s worldwide grocery team, which includes Whole Foods.
Rather than just making a larger version of the consumer bin for commercial use, Mill realized that it could go a step farther. By adding cameras to the bin and using AI, it’s possible to identify the food scraps so a restaurant or grocery kitchen can get a detailed report of exactly what’s being thrown out. The team showed me how it worked on a test rig where they’re calibrating the computer vision system.
After someone dumped in a bucket of scraps, I watched the results show up on a computer screen—1.5 pounds of watermelon rinds, half a pound of potatoes, half a pound of grapes, and a metal fork. (Flatware is surprisingly common in restaurant trash; the new system flags it so it can be taken out and staff can be trained to stop dumping it.)
“We’re able to characterize the kind of food that’s thrown away in real time,” says Tannenbaum. “So think down to item level visibility into exactly what’s being thrown into a Mill. If you know exactly what you’re wasting, then you can attack it and triage it and it becomes much more tractable.”
The program is designed to connect to other software that kitchens use, such as meal planning programs. Eventually, managers could choose to let AI run some of the process; if a portion of a particular dish is too big, Mill’s tech can recommend how much to adjust the size and then connect with other programs to order fewer ingredients, notify the head chef, and adjust prep based on the menu change.
The company isn’t the only one to design technology to help kitchens track what’s being thrown out. But other options require more steps—for example, holding food in front of a static camera for a few seconds, then putting it on a scale, and then using a touchscreen to confirm what the item was.
“We’ve heard from customers it’s too fragmented,” says Kristen Verdone, head of product at Mill. “It’s too hard to use, and it takes too much time. When that’s the case, then staff doesn’t use it and then they don’t get the data that they need.”
Mill knew that its product had to be as easy to use—or easier—than dumping food in the trash. To get there, the team has spent months researching what actually happens in kitchens to make sure they get the user experience right.
“The kitchen environment is incredibly hectic,” says Irina Kozlovskaya, Mill’s industrial design lead. “The pace is super fast, it’s hot, it’s humid, it’s really, really intense in there. What we saw actually confirmed our hypothesis: tasks that can be skipped really get skipped. This again put extra emphasis on how important it was for us to make this device really easy to use, and catering to that end user need.”
The process means spending a lot of time around trash, everywhere from university cafeterias to hospital kitchens to cruise ships. At one memorable point at the beginning of the research, Verdone stood in 100-degree heat on a dock behind a large hotel kitchen chewing gum to keep from gagging at the smell of rotting garbage in the trash compactor.
“[That was] my first real look at the contrast between pristine front-of-house dining and what happens behind the scenes,” she says. It was clear that Mill’s system could help remove some pain points for staff—the smell, the pests attracted by garbage, and taking frequent trips to haul heavy, dripping trash to the back.
They experimented with different sizes of the device to find the right balance between holding the biggest volume of food scraps possible while still being small enough to fit in a crowded kitchen—roughly the size of a dishwasher—so staff don’t have to go out of their way to use it. The angle of the bin’s opening is ergonomically designed to make it as easy as possible for someone to empty a heavy container of food scraps. The UX is designed to be intuitive, so someone can walk up to the bin and use it without training.
In one corner of Mill’s R&D facility, the design team is tweaking the final shape of the device. Across the room, engineers are working on a non-form factor prototype with all of the components exposed. The process is the same as the team used at companies like Apple. It “starts with this parallel track of what’s the ideal dream state that meets all those user touchpoint needs, and what are the engineering functional bits that need to come together,” says Adam Middleman, who leads the engineering team and previously worked at Apple. “We start on these parallel tracks and as we get closer and closer to production, they come together into the same thing.”
Nearby, other engineers are working on the computer vision system that identifies the food, and on clear software to make the results usable. “We’ve learned from our customers that they’re tired of spreadsheets—there’s too much data, they just want to know what they need to do,” says Azita Sayadi, Mill’s commercial product lead. “There’s limited time. The metric that matters most is savings, because cost is really important as it relates to food waste.”
The screen shows the top five items that are being wasted, and how much could be saved with a recommended action like retesting a recipe or shrinking a portion size.
After more R&D, the new device will start rolling out to customers next year at a yet-to-be-announced price. Some, like Whole Foods, plan to use the dehydrated food waste in their supply chains; Whole Foods is talking to its egg suppliers about incorporating the material in chicken feed. Others may use the system to make it more affordable to ship food waste for composting or for use in fertilizer.
At the same time, the system has the potential to significantly shrink the amount of waste that goes out the door. It’s fair to say that it’s a problem that hasn’t gotten attention from former Apple engineers in the past. “This is one of those problem spaces that desperately needs a team like ours,” says Rogers. “About 10% of global emissions is food waste. About $400 billion of wasted food in the U.S. per year. But there certainly were not the crack A teams from Apple and Nest working on them.”